Induction of labor before 42 completed weeks of gestation increased steadily in the United States between 1990 and 2010.1,2 This increase reflects rises in rates of induction with and without medical indication (also known as elective induction of labor). However, the evidence about nonmedically indicated induction of labor and its effect on a variety of maternal and neonatal outcomes is not clear. Data supporting induction of labor for women at 41 weeks of gestation and beyond exist,3 but less is known about the effect of induction without medical indication between 37 and 40 completed weeks of gestation.3 With retrospective evidence indicating that early-term (ie, 37 and 38 weeks of gestation) delivery confers higher risk for subsequent adverse neonatal and childhood outcomes compared with later-term weeks,4 the American College of Obstetricians and Gynecologists has issued recommendations to reduce nonmedically indicated induction of labor at less than 39 weeks of gestation.5 Recent evaluations of strategies to reduce induction in the absence of medical indication before 39 weeks of gestation have reported decreases in admissions to the neonatal intensive care unit (NICU),6,7 conflicting results about stillbirth,7,8 and little information about cesarean delivery, historically one of the key concerns surrounding induction without medical indication.9
The lack of transparent, reproducible methods to classify inductions as medically indicated or not and to define appropriate comparison groups is a key contributor to the evidence gap about the health effects of induction of labor without medical indication. This analysis focuses on induction of labor without medical indication and expectant management at each term gestational week (37–40 weeks). We improve on prior work by using a transparent method to classify inductions as nonmedically indicated and the clinically relevant comparison group, expectant management.10 We stratify by gestational age and parity and test the association of induction without a medical indication and cesarean delivery, operative vaginal delivery, third- or fourth-degree perineal lacerations, perinatal death, NICU admission, respiratory distress, hyperbilirubinemia, shoulder dystocia, and macrosomia.
MATERIALS AND METHODS
We conducted a retrospective cohort study using 2006 California Department of Health Services linked data (death files, birth certificates, and unmasked hospital discharge data).11 It contains linked birth and delivery records that contain deidentified information for a mother and neonate pair from neonatal and maternal discharge data and the birth certificate data (N=532,088) and includes all deliveries in a given year.
We arrived at our analytic sample of 362,154 after a series of exclusions (Fig. 1). In the induction without medical indication group, we included women who delivered between 37 and 40 completed weeks of gestation because late-term or postterm pregnancy (greater than 41 or 42 completed weeks of gestation) is a common indication for induction, and good evidence already exists to support induction for such pregnancies.3,10 We used The Joint Commission list of indications possibly justifying delivery before 39 completed weeks of gestation12 to classify women with induction without medical indication (see Appendix 1, available online at http://links.lww.com/AOG/A431).
A limitation of previous observational studies is choice of comparison group: women undergoing induction have been compared with women who had spontaneous labor in that same week of gestation.9,13–16 However, the clinical reality is not a choice between induction or spontaneous labor but between induction and continuing pregnancy, ie, expectant management with the potential for either spontaneous or induced labor and delivery at a later gestational age.17 To ensure that our findings reflected clinical practice, we compared women with nonmedically indicated inductions at each term week of gestation with expectant management (a group of women who deliver in the next week or beyond, up to 42 completed weeks of gestation) (Fig. 1).
We further divided the sample by gestational age such that all women with nonmedically indicated inductions at 37 completed weeks of gestation were compared with women who delivered at 38–42 completed weeks of gestation, excluding women at each week with deliveries who had antepartum indications (Fig. 1). At each week of induction without a medical indication, the comparison group shifts to include only those women delivered after the week of induction. Because this method assumes that expectant management continues into the next week, we also performed a sensitivity analysis comparing induction without a medical indication at each week with women at that week or above.18,19
We used hospital discharge International Classification of Diseases, 9th Revision, Clinical Modification diagnosis and procedure codes to identify cesarean delivery (669.7, 669.70, 669.71, 763.4, 74, 74.0, 74.1, 74.2, 74.4, 74.9, 74.99), third-or fourth-degree perineal lacerations (664.2, 664.20, 664.21, 664.24), operative vaginal delivery (669.5, 669.50, 669.51, 72.8, 72.9), respiratory distress (769, 770, 770.89, 770.84, 770.9), hyperbilirubinemia (277.5, 773.4, 774, 774.0, 774.1, 774.2, 774.3, 774.30, 774.31, 774.39, 774.5, 774.6, 774.7, 782.4), and shoulder dystocia (660.4, 660.40, 660.41, 660.43). Neonatal intensive care unit admission was a composite of NICU admission or transfer documented on the birth certificate and neonatal length of stay greater than maternal length of stay documented in the discharge record. Perinatal death was a composite of fetal, neonatal, and infant death obtained from vital statistics files. Macrosomia (birth weight greater than 4,000 g) was recorded on the birth certificate. Covariates extracted from hospital discharge or birth certificate data included advanced maternal age (35 years or older) at the time of delivery, initiation of prenatal care in the first trimester, insurance status (none, public, private), maternal education (completed high school or not), maternal race, and delivery at a teaching hospital. We did not control for birth weight because increasing birth weight is an inherent risk of expectant management and is also on the causal pathway from expectant management to several neonatal outcomes including macrosomia. Some covariates are available in both the birth certificate and in the hospital discharge data. We privileged the hospital discharge data over the birth certificate information when possible, congruent with data validation studies.20–22
We compared proportions of each binary outcome between induction without a medical indication and expectant groups stratified by parity (nulliparous women or previous vaginal delivery) using a two-sample test of proportions. We used multivariable logistic regression models to estimate adjusted associations between induction without a medical indication and our predetermined outcomes. We used robust standard errors to account for data clustering at the hospital level.23 We included common individual and institutional-level confounders (listed previously) in our models. We included only women with vaginal deliveries in our model for shoulder dystocia. We anticipated that cesarean delivery would be a common outcome and therefore that odds ratios (ORs) would overstate associations if interpreted as relative risks, so we also estimated relative risks24 but estimates were similar so we present only the OR estimates.
Our elective induction classification scheme assumes that all documented indications were known before the decision to induce. Our data do not permit us to assess temporality, but it is plausible that some conditions could develop after an induction without medical indication. For example, fetal distress (656.31) could occur before the decision to induce, in which case it is an indication for induction, or after the induction, in which case it is part of the risk of an elective induction. We therefore changed our assumptions in a sensitivity analysis where we retained intrapartum conditions that could plausibly follow elective induction in the elective induction group (see Appendix 1, http://links.lww.com/AOG/A431). All analyses were conducted using STATA 12. This study was approved by the California Office of Statewide Health Planning and Development, Oregon Health & Science University institutional review board, and the University of California, San Francisco Committee on Human Research.
The analytic sample included 362,154 term vertex singleton deliveries (46.5% nulliparas). Overall, 5.4% of all deliveries were identified as induced without a documented indication (elective), whereas 11.4% of included deliveries were identified as induced with a documented indication. At each term week of gestation, a greater proportion of women who were induced without medical indication were white, had private insurance, had completed high school, and had initiated prenatal care in the first trimester compared with expectant management (Table 1). A smaller proportion of women who were induced without medical indication delivered at a teaching hospital or were nulliparous compared with expectant management. Overall, the cesarean delivery rate was 16% (n=58,667) (26.5% among nulliparous women, 7.3% among multiparous women); perinatal mortality was 0.2% (n=706); and NICU admission, transfer, or a neonatal hospital stay greater than that of the mother was 6.2% (n=22,409). In bivariate analyses (Table 2), the proportion of cesarean deliveries was significantly larger at each gestational age in the expectant management groups (37 weeks of gestation 15.4% compared with 7.6%; 38 weeks of gestation 15.9% compared with 8.0%; 39 weeks of gestation 17.3% compared with 9.3%; 40 weeks of gestation 19.0% compared with 12.4%). The proportion of hyperbilirubinemia and shoulder dystocia in the elective induction groups was higher than in the expectant groups at 37 and 38 weeks and at 39 weeks of gestation, respectively. There was either no difference or a lower proportion of all other outcomes in the induction without medical indication groups than in the expectant management groups in bivariate analyses (Table 2).
The odds of cesarean delivery were significantly lower among women in the induction without medical indication group at 37 completed weeks of gestation (OR 0.44, 95% confidence interval [CI] 0.34–0.57), 38 weeks of gestation (OR 0.43, CI 0.38–0.50), 39 weeks of gestation (OR 0.46, CI 0.41–0.52), and 40 weeks of gestation (OR 0.57, CI 0.50–0.65) (Table 3). Although this relationship was especially strong among multiparous women, it held among nulliparous women at each week. We found reduced odds of operative vaginal deliveries among women induced without medical indication in the full sample but no differences between induction without a medical indication and expectant management in the odds of operative vaginal deliveries in analyses stratified by parity except at 40 weeks of gestation among multiparous women (OR 0.78, CI 0.60–0.99). We show similar results for third- or fourth-degree perineal lacerations.
Elective induction of labor was not associated with significantly higher odds of perinatal death, NICU admission, or respiratory distress at any gestational age or parity (we were unable to stratify by parity for perinatal death). However, we cannot rule out increased odds of NICU admission or respiratory distress at 37 and 38 weeks of gestation and shoulder dystocia at 37, 38, 39, and 40 weeks of gestation given wide CIs on these outcomes (Table 4). Elective induction was associated with increased odds of hyperbilirubinemia at 37 (OR 1.29, CI 1.05–1.59) and 38 (OR 1.17, CI 1.00–1.37) weeks of gestation among multiparous women and at 38 (OR 1.27, CI 1.06–1.53) weeks of gestation among nulliparous women and we cannot rule out increased risk at every gestation age (Table 4). Women undergoing induction without a medical indication had reduced odds of delivering a neonate with macrosomia up to 40 weeks of gestation, when there was no difference by group.
In sensitivity analyses, changing our classification scheme to include deliveries “at or above”18 the week of induction did not alter our results (data not shown). Changing assumptions about the temporality of intrapartum indications did alter results for cesarean delivery, respiratory distress, and hyperbilirubinemia outcomes. Under this scenario, the elective induction group either moved toward the null (for cesarean delivery, respiratory distress) or showed increased risk (hyperbilirubinemia) (Appendix 2, available online at http://links.lww.com/AOG/A431).
In our analysis of women with term, singleton, vertex pregnancies in California, we found that induction of labor without medical indication was associated with reduced odds of cesarean delivery among both nulliparous and multiparous women at each term gestational age (37–40 weeks). We conducted sensitivity analyses, which included women in the elective induction group who had plausible intrapartum complications, rendering the induction group higher risk, and found that induction without medical indication was not associated with increased odds of cesarean delivery at any term gestational age and was still reduced, although attenuated, among multiparous women. In addition, we found that elective induction was not associated with increased perinatal mortality or NICU admission at any term gestational age (37–40 weeks) compared with expectant management.
This analysis extends previous observational studies of elective induction by including all term gestational weeks (37–40 weeks), stratifying by parity, and using a transparent method to classify elective induction and expectant management groups. Labor induction without a medical indication is complicated to study in observational data, and our methods improve on previous approaches. Evidence is mounting that elective induction does not increase risks of cesarean delivery and may actually reduce risks of certain outcomes; however, results depend on the analytical method used. Our findings generally support earlier studies that used expectant management comparison groups19,25 despite differences in classifying induction cases, data sources, and exclusion criteria. Recently, there have been numerous attempts to reduce induction of labor without medical indication before 39 weeks of gestation and in some cases to reduce induction without an indication overall. Unfortunately, these efforts are based on a relatively limited literature and like many medical decisions, there are likely tradeoffs in the use of induction of labor without medical indication.26 For example, Ehrenthal and colleagues7 reported that although NICU rates declined after an institutional policy change to restrict elective deliveries at less than 39 completed weeks of gestation, rates of macrosomia and stillbirth increased.
Our sensitivity analyses, which varied assumptions about the timing of indications and thus about the relative risks inherent in expectant management and induction without prior indication, demonstrate the importance of temporality assumptions in retrospective analyses. Our main analysis is conceptually sound, assuming that intrapartum conditions are an inherent risk of expectant management. Our data, however, do not provide temporality of complications that would most commonly occur intrapartum (eg, chorioamnionitis, fetal heart rate abnormalities). Furthermore, the need for a consensus set of medical indications and associated International Classification of Diseases, 9th Revision, Clinical Modification codes for induction will become more pressing as research continues to focus on nonmedically indicated deliveries as a quality metric.27,28 Prospective data collected for research purposes are necessary to arrive at a conclusive answer, but this analysis points to potential reasons that prior prospective and retrospective work has been so divergent.
This analysis is subject to the limitations of all observational studies. We rely on vital statistics and hospital discharge data sources, which were not designed to answer our study question and likely lead to some misclassification. Gestational age dating using last menstrual period is subject to error but has shown high concordance with clinical estimate dating at term (37–41 weeks of gestation).29 International Classification of Diseases, 9th Revision, Clinical Modification codes, although better than vital statistics alone to study obstetric management,20 have varying validity for both induction and medical indications. We have likely missed some inductions (if the procedure was not recorded in the hospital discharge data) and also misclassified some inductions as nonmedically indicated (if medical indications were not recorded) or as medically indicated (if indications were recorded in error). This could bias our results to favor elective induction (if more women are misclassified from elective induction to indicated induction than from indicated to elective). We have controlled for measured variables in our multivariable analyses, but we recognize that there could be additional unmeasured confounders (eg, patient preferences and other health behaviors) and that the distribution of characteristics we were able to measure suggests possible “healthy user” bias (eg, induced women have higher education and earlier prenatal care, which may go along with other health-seeking behaviors). We used The Joint Commission list of indications and codes; although this definition may misclassify some women it is a standard for reporting, and we need to understand how it shapes classification and estimation of outcomes. Although we have a large data set, we report wide uncertainty estimates for rare outcomes such as perinatal mortality and may require a larger sample to compute a more precise estimate. We account for data clustering at the hospital level but do not have information about usual care at each hospital, which could inform observed induction or cesarean delivery rates. We have no data on cervical status, which has been found to be important in the relationship of induction without medical indication and cesarean delivery, although it has very limited ability to predict whether induction will result in a vaginal delivery.30 Finally, we were unable to identify planned and unplanned repeat cesarean delivery and thus could not include this group in our analysis.
In conclusion, we present evidence that induction without medical indication at term (37–40 weeks of gestation) is associated with reduced odds of cesarean delivery among both nulliparous and multiparous women with a previous vaginal delivery. This holds for multiparous women at 38 and 39 completed weeks of gestation even when we vary assumptions about the timing of intrapartum indications. With the exception of hyperbilirubinemia at early-term gestational ages, we find no evidence of any other increased adverse maternal or neonatal outcomes with elective induction. Focus on induction of labor as a quality metric in obstetrics must be evidence-based. The use of a standard method to classify induction without medical indication and use of the appropriate comparison groups by researchers would permit comparison across studies and improve our ability to draw conclusions about the effect of elective induction on maternal and neonatal health.
1. Chauhan SP, Ananth CV. Induction of labor in the United States: a critical appraisal of appropriateness and reducibility. Semin Perinatol 2012;36:336–43.
2. Murthy K, Grobman WA, Lee TA, Holl JL. Trends in induction of labor at early-term gestation. Am J Obstet Gynecol 2011;204:435.e1–6.
3. Gulmezoglu AM, Crowther CA, Middleton P, Heatley E. Induction of labour for improving birth outcomes for women at or beyond term. The Cochrane Database of Systematic Reviews 2012, Issue 6. Art. No.: CD004945. DOI: 10.1002/14651858.CD004945.pub3.
4. Reddy UM, Bettegowda VR, Dias T, Yamada-Kushnir T, Ko C-W, Willinger M. Term pregnancy: a period of heterogeneous risk for infant mortality. Obstet Gynecol 2011;117:1279–87.
5. Induction of labor. ACOG Practice Bulletin No. 107. American College of Obstetricians and Gynecologists. Obstet Gynecol 2009;114:386–97.
6. Clark SL, Frye DR, Meyers JA, Belfort MA, Dildy GA, Kofford S, et al.. Reduction in elective delivery at <39 weeks of gestation: comparative effectiveness of 3 approaches to change and the impact on neonatal intensive care admission and stillbirth. Am J Obstet Gynecol 2010;203:449.e1–6.
7. Ehrenthal DB, Hoffman MK, Jiang X, Ostrum G. Neonatal outcomes after implementation of guidelines limiting elective delivery before 39 weeks of gestation. Obstet Gynecol 2011;118:1047–55.
8. Oshiro BT, Henry E, Wilson J, Branch W, Varner MW; Women and Newborn Clinical Integration Program. Decreasing elective deliveries before 39 weeks of gestation in an integrated health care system. Obstet Gynecol 2009;113:804–11.
9. Vahratain A, Zhang J, Troendle JF, Sciscione AC, Hoffman MK. Labor progression and risk of cesarean delivery in electively induced nulliparas. Obstet Gynecol 2005;105:698–704.
10. Caughey AB, Sundaram V, Kaimal AJ, Gienger A, Cheng YW, Mcdonald KM, et al.. Systematic review: elective induction of labor versus expectant management of pregnancy. Ann Intern Med 2009;151:252–63, W53–63.
11. California Department of Health Services. Center for Health Statistics. Birth cohort public use file, 1999-2003. In: California Department of Health Services, editor. Sacramento (CA): California Department of Health Services; 2006.
13. Seyb ST, Berka RJ, Sccol ML, Dooley SL. Risk of cesarean delivery with elective induction of labor at term in nulliparous women. Obstet Gynecol 1999;94:600–7.
14. Jonsson M, Cnattingius S, Wikström AK. Elective induction of labor and the risk of cesarean section in low-risk parous women: a cohort study. Acta Obstet Gynecol Scand 2013;92:198–203.
15. Cammu H, Martens G, Ruyssinck G, Amy JJ. Outcome after elective labor induction in nulliparous women: a matched cohort study. Am J Obstet Gynecol 2002;186:240–4.
16. Hoffman MK, Vahratian A, Sciscione AC, Troendle JF, Zhang J. Comparison of labor progression between induced and noninduced multiparous women. Obstet Gynecol 2006;107:1029–34.
17. Caughey AB, Nicholson JM, Cheng YW, Lyell DJ, Washington AE. Induction of labor and cesarean delivery by gestational age. Am J Obstet Gynecol 2006;195:700–5.
18. Glantz JC. Term labor induction compared with expectant management. Obstet Gynecol 2010;115:70–6.
19. Stock SJ, Ferguson E, Duffy A, Ford I, Chalmers J, Norman JE. Outcomes of elective induction of labor compared with expectant management: population based study. BMJ 2012;344:e2838.
20. Goff SL, Pekow PS, Markenson G, Knee A, Chasan-Taber L, Lindenauer PK. Vailidity of using ICD-9-CM codes to identify selected categories of obstetric complications, procedures and co-morbidities. Paediatr Perinat Epidemiol 2012;26:421–9.
21. Lain SJ, Hadfield RM, Raynes-Greenow CH, Ford JB, Mealing NM, Algert C, et al.. Quality of data in perinatal population health databases. Med Care 2012;50:e7–20.
22. Lyndon-Rochelle MT, Holt VL, Cardenas V, Nelson JC, Easterling TR, Gardella C, et al.. The reporting of pre-existing maternal medical conditions and complications of pregnancy on birth certificates and in hospital discharge data. Am J Obstet Gynecol 2005;193:125–34.
23. Diggle PJ, Heagerty P, Liang K-Y, Zeger SL. Analysis of longitudinal data. New York, (NY): Oxford University Press; 2002.
24. McNutt L-A, Wu C, Xue X, Hafner JP. Estimating the relative risk in cohort studies and clinical trials of common outcomes. Am J Epidemiol 2003;157:940–3.
25. Cheng YW, Kaimal AJ, Snowden JM, Nicholson JM, Caughey AB. Induction of labor compared to expectant management in low-risk women and associated perinatal outcomes. Am J Obstet Gynecol 2012;207:502.e1–8.
26. Macones GA. Elective induction of labor: waking the sleeping dogma? Ann Intern Med 2009;151:281–2.
27. Main EK, Morton CH, Melsop K, Hopkins D, Guiliani G, Gould JB. Creating a public agenda for maternity safety and quality in cesarean delivery. Obstet Gynecol 2012;120:1194–8.
28. Spong CY, Berghella V, Wenstrom KD, Mercer BM, Saade GR. Preventing the first cesarean delivery: summary of a joint Eunice Kennedy Shriver
National Institute of Child Health and Human Development, Society for Maternal-Fetal Medicine, and American College of Obstetricians and Gynecologists Workshop. Obstet Gynecol 2012;120:1181–93.
29. Ananth CV. Menstrual versus clinical estimate of gestational age dating in the United States: temporal trends and variability in indices of perinatal outcomes. Paediatr Perinat Epidemiol 2007;21(suppl 2):22–30.
© 2013 by The American College of Obstetricians and Gynecologists. Published by Wolters Kluwer Health, Inc. All rights reserved.
30. Grobman WA. Predictors of induction success. Semin Perinatol 2012;36:344–7.